近年来，游戏化理念已经在多个领域进行了成功的实践并受到了很多重视，但这一理念在人事测评领域，尤其是和评价中心技术相结合的应用研究仍然较少。采用游戏这一载体进行工作情景相关研究的理念，看似与传统认知中“游戏和工作完全不兼容”的观点相悖，实际上游戏化技术在管理、教育以及认知测验上的应用成效是十分显著的。公文筐测验是评价中心最常用并且最核心的技术之一，在国外公文筐测验的使用频率高达81%。虽然该测验早就被证实为是一种有效的人才选拔和评估方式，但这一测验形式与游戏化的结合十分少见。综上所述，本研究旨在利用游戏化公文筐测验的形式，开发基于游戏形式的、问题情境的、适用于企业在职员工的测验，考察沟通、计划组织以及决策这三种能力。本研究共包含两个部分，具体内容如下：

研究一是基于游戏化的公文筐测验开发。首先，对有工作经验的在职员工进行行为访谈，提取工作中的典型情境，为兼顾情境典型性和连贯性，在开发过程中将电话录音表、微信消息、邮件、匿名投诉信、便签等5种形式的6个公文筐测验编制在了同一故事背景下。其次，确定公文筐的答题形式、题目结构、数量以及组合方式，通过专家评定与修改后形成书面版的公文筐测验和题目。最后进行测验的游戏化，包括游戏的逻辑设计、元素设计以及UI布局，通过寻求专业游戏开发团队的帮助，经过多次修改形成最终的游戏化公文筐测验。

研究二是对公文筐测验的题目质量和效度进行检验。本研究使用游戏平台，采用线上的形式对188名在职人员进行测试。同时收集了人际沟通能力问卷、决策能力量表以及计划组织能力量表数据，用以检验游戏化公文筐测验的效度。根据项目反应理论的结果显示，该测验在有工作经验的在职员工样本中具有良好的难度和区分度。此外相容效度结果表明，游戏中的沟通能力与人际沟通量表呈显著正相关，而决策能力和计划组织能力的结果并不理想。除此之外，本研究尝试通过BP神经网络探讨游戏化公文筐测验中任务作答顺序对计划组织能力和决策能力的影响，以检验任务作答顺序对个体能力的预测效果。由于样本数据量的不足，导致BP神经网络的预测结果并不理想，在讨论部分探讨了可能导致该结果的原因，以及今后使用该预测模型时的改进建议。

In recent years, the concept of gamification has been successfully practiced in many fields and received more and more attention. However, there is still little research on its application in the field of personnel evaluation, especially the combination of the technology with Assessment Center. The idea of using games as a carrier for research seems to contradict the traditional view that "games and work are completely incompatible", but it is undeniable that it has been applied to management, education and cognitive tests with remarkable results. In-Basket Test is commonly used and core technologies of the Assessment Center. The frequency of utilization of the In-Basket Test in foreign countries has reached 81%. As an effective way of talent selection and evaluation, the In-Basket Test is rarely combined with gamification. The purpose of this research is to develop a game-based, problem-based, In-Basket Test in the form of gamification for employees of in-service companies, and to examine three abilities including communication skills, planning and organization skills, and decision-making skills. This research consists of two parts, details are as follows:

Study one is the development of In-Basket Test based on game. Firstly, through behavioral interviews with experienced employees, typical situations at work were extracted. To take into account of both the typicality and coherence of situations, 6 In-Basket Tests in 5 forms, such as telephone recording form, WeChat message, email, anonymous complaint letter and note, were compiled under the same story background. Next, the answer form, topic structure, quantity and combination of the in-basket were determined, then evaluated and modified by experts. Finally formed into a written version of the In-Basket Test and questions. About the gamification development, first things carried out were the logical design, element design and UI layout of the game. Then, a professional development team was found to form the final gamification In-Basket Test after several modifications.

Study two is to test the quality and validity of the In-Basket Test. 188 working people were tested online through a gamification platform. At the same time, data from the Interpersonal Communication Skills Questionnaire, Decision Making Skills Scale and Planning Organization Scale were collected to test the validity of the gamified in-basket test. According to the results of item response theory, this test has a good degree of difficulty and differentiation in the sample of employees with working experience. The results of criterion correlation validity showed that the communication ability in the game was positively correlated with the interpersonal communication scale, while the results of decision-making ability and planning and organizing ability were not ideal. In addition, BP neural network was used to explore the influence of task order on planning and organization ability and decision-making ability in the gamification in-basket test, as to test the prediction effect of task order on individual ability. However, due to the lack of sample data, the prediction result of BP neural network is not ideal. In the discussion section, the possible reasons for this result are and the suggestions for improvement when using the prediction model in the future are discussed.